A Hybrid Approach using Fuzzy Logic and Neural Network for Enhancement of Low Contrast Images

Transcription

1 A Hybrid Approach using Fuzzy Logic and Neural Network for Enhancement of Low Contrast Images Ms. Manu Gupta M.Tech Student, RIET Phagwara Er. Amanpreet Kaur Chela AP, CSE, RIET Phagwara Abstract: Many computer vision or machine vision researchers are working on the field of image enhancement now days. The main attraction towards this research area is because of the additional knowledge and hidden information provided by the results of this procedure which will further be used for many different useful purposes. Number of algorithms were purposed for enhancing an image like Histogram Equalization, High boost filtering etc. The different results which we can get from these methods can be compare against the parameters like MSE (Mean Square Error), RMSE (Root Mean Square Error), SNR (Signal to Noise Ratio), PSNR (Peak Signal to Noise Ratio). In this paper we proposed a new hybrid technique by considering two optimization techniques Artificial Neural Network (ANN) and Fuzzy Logic for image enhancement and want to compare results with the existing techniques by considering above said parameters. Keywords: Fuzzy Logic, ANN, Histogram Equalization, High boost filtering, Computer Vision. I. INTRODUCTION An image may be defined as a two dimensional function f(x, y), where x and y are spatial (plane) coordinates and the amplitude of f at any pair of coordinates (x, y) is called the intensity or gray level of the image at that point. When x, y and the amplitude values are all finite, discrete quantities, the image is called a digital image. A digital image is a representation of a two-dimensional image as a finite set of digital values. It is composed of a finite number of elements each of which has a particular location and value. These elements are referred to as picture elements, image elements and pixels. Image enhancement processes consist of a collection of techniques that seek to improve the visual appearance of an image or to convert the image to a form better suited for analysis by a human or a machine. Figure1: Enhancement Process Image enhancement techniques can be divided into three broad categories: 1) Spatial domain methods, which operate directly on pixels. 2) Frequency domain methods, which operate on the Fourier transform of an image. 3) Fuzzy domain methods, which involves the use of knowledge-base systems. Whenever an image is converted from one form to another such as digitizing, scanning, transmitting, storing, etc. some degradation occurs at the output. Hence, the output image has to undergo a process called image enhancement. Fuzzy image processing is the collection of all approaches that understand, represent and process the images, their segments and features as fuzzy sets. The representation and processing depend on the selected fuzzy technique and on the problem to be solved. The idea of fuzzy sets is simple, natural and on the basis for human communication. Because fuzzy logic is built on the structures of qualitative description used in everyday language, fuzzy logic is easy to use. A filtering system needs to be capable of reasoning with vague and uncertain information. This suggests the use of fuzzy logic. There is no general theory for determining what good image enhancement is when it comes to human perception. If it looks good, it is good! However when image enhancement techniques are used as pre-processing tools for other image processing techniques, then quantitative measures can determine which techniques is most appropriate. Fuzzy image enhancement is based on gray level mapping into a fuzzy plane, using a membership transformation Function. The aim is to generate an image of higher contrast than the original image. An image I of size M x N and L gray level can be considered as an array of fuzzy singletons, each having a value of membership denoting its Input Image Enhancement Output Image Degree of brightness relative to some brightness levels

2 International Journal of Scientific Research Engineering & Technology (IJSRET) Input Image Image Fuzzification Membership modification Defuzzification Enhanced image Figure2: Fuzzy Logic Implementation The fuzzification and de-fuzzification steps are due to the fact that we do not possess fuzzy hardware. Therefore, the coding of image data (fuzzification) and decoding of the results (de - fuzzification) are steps that make possible to process images with fuzzy techniques. The main power of fuzzy image processing is in the middle step (modification of membership values). After the image data is transformed from gray- plane level plane to the membership (fuzzification), appropriate fuzzy techniques modify the membership values. This can be a fuzzy clustering, a fuzzy rule-based approach, and a fuzzy integration approach and so on. If we interpret the image features as linguistic variables, then we can use fuzzy if-then rules to segment the image into different regions. A simple fuzzy segmentation rule may seem as follows: IF the pixel is dark AND its neighborhood is also dark AND homogeneous THEN it belongs to the background. Figure3: Example of Fuzzification and De-fuzzification An Artificial Neural Network (ANN) is an information-processing paradigm that is inspired by the way a biological nervous system in human brain works. Large number of neurons present in the human brain forms the key element of the neural network paradigm and act as elementary processing elements. These neurons are highly interconnected and work in unison to solve complex problems. Likewise, an Artificial Neural Network can be

3 configured to solve a number of difficult and complex problems. ANNs find a wide variety of applications in diverse areas including functional approximation, nonlinear system identification and control, pattern recognition and pattern classification, optimization, English text pronunciation, protein secondary structure prediction and speech recognition. The biological neurons have many useful and significant characteristics and properties. These are also emulated by neurons in artificial neural networks. Some of these features of artificial neural networks are outlined as under: Input-output Mapping Learning With Experience Nonlinearity Model Free Environment Hardware Implementation Parallel Distributed Processing Multivariable Systems Fault Tolerance Data Fusion There are many reasons to use fuzzy and Neural Network technique. The most important of them are as follows: 1) Fuzzy techniques are powerful tools for knowledge representation and processing. 2) Fuzzy techniques can manage the vagueness and ambiguity efficiently. 3) Neural Network is consider as an Optimization technique and helps for making decisions in accordance with the weights of Neurons. Enhancement of noisy image data is a very challenging issue in many research and application areas. When the images are transmitted over channels, they are corrupted with impulse noise due to noisy channels. This impulse noise consists of large positive and negative spikes. The positive spikes have values much larger than the background and thus they appear as bright spots, while the negative spikes have values smaller than the background and they appear as darker spots. Both the spots for the positive and negative spikes are visible to the human eye. Also, Gaussian type of noise affects the image. Thus, filtering for both impulse noise and Gaussian noise is required before processing. There are lots of classical and fuzzy filters and Neural Network filters in the literature to remove noise. The classical filters are the mean filter, the Gaussian filter, the median filter, the adaptive median filter, high boost filtering etc. The mean filter or the average filter helps in smoothing operations. It suppresses the noise that is smaller in size or any other small fluctuations in the image. It involves in calculating the average brightness values in some neighborhood m x n and replaces the gray level with an average value. Smoothing or averaging operation blurs the image and does not preserve the edges. These are not useful in removing noise spikes. Fuzzy and Artificial Neural Network filters provide promising result in image-processing tasks that cope with some drawbacks of classical filters. Fuzzy filter is capable of dealing with vague and uncertain information. Sometimes, it is required to recover a heavily noise corrupted image where a lot of uncertainties are present and in this case fuzzy set theory is very useful. Each pixel in the image is represented by a membership function and different types of fuzzy rules that considers the neighborhood information or other information, classical filters removes the noise with blurry edges but fuzzy filters perform both the edge preservation and smoothing. Thus, the combination of fuzzy logic and Neural Network can be used together for the enhancement of low contrast images corrupted with noise. II. OBJECTIVE OF PROPOSED WORK 1) Image Preprocessing 2) Analysis of the existing enhancement methods in spatial, frequency domain and Fuzzy domain methods. 3) Design the technique using FIS ( Fuzzy Inference System) to improve the image contrast in the presence of Neural Network. 4) An algorithm is proposed and implemented to enhance images using fuzzy logic and Neural Network technique. This algorithm is used to convert the image properties into fuzzy data and fuzzy data into de-fuzzification and in between a Neural Network is generated to identify edge pixels for improving enhancement process. 5) Test the designed technique using some degraded, low contrast images. 6) The designed technique is able to improve contrast of the image as compare to Histogram Equalization, Spatial Averaging Filter, Median Filter, Un-Sharp Masking & High Boost filtering methods. III. METHODOLOGY OF PROPOSED WORK In the implementation part firstly we took a low contrast image or we can input a normal image and then add Gaussian noise to that image to check the performance of our proposed method on noisy image.

4 Then we implement the existing techniques on the noisy image and got some enhanced images then we follow the proposed process of hybrid technique where we use fuzzy logic and neural network for image enhancement here firstly we use the concept of neural network where we apply 14 different fuzzy rules on the noisy image and prepare the weights of Neural network. Resize Gray Scale Conversion Addition of Noise Output Image of Phase-I Histogram Equalization Spatial Averaging Filter Median Filters Un-Sharp Masking & High Boost Fuzzy Logic & Neural Network System Enhanced Image1 Enhanced Image2 Enhanced Image3 Enhanced Image4 Enhanced Image5 Evaluation of enhanced image results Figure4: Proposed Methodology And then from those weights, we can find some decisive values which then used in fuzzification process. Here we proposed three membership functions for fuzzification process as one triangular membership function and two trapezoidal membership functions. Then we apply the defuzzification process to get the enhanced image after the entire processing. At the end we can compare different enhanced images of existing techniques with the enhanced image got from the proposed method. The comparison is done on the basis of parameters MSE, RMSE, SNR and PSNR. if the final image will have the values of MSE and RMSE low as compared to other and value of SNR and PSNR high then this will show the quality of the output image

5 IV. CONCLUSION Existing techniques of image enhancement are not effective in dealing with imprecise data, vague and uncertain information. Most of the existing techniques are either very sensitive to noise and do not give satisfactory results in low contrast and light variations areas. Noise smoothing and edge enhancement are inherently conflicting processes, since smoothing a region might destroy an edge, while sharpening edges might lead to unnecessary noise. There is requirement of technique which can enhance the images while edge preservation, noise removal and smoothing. Because of the uncertainty that exist in many aspects of image processing, fuzzy processing is desirable. These uncertainties include additive and non-additive noise in low level image processing, imprecision & ambiguities in interpretation during high level image processing. Number of decisions have to make to decide which pixel can be consider as Edge pixel or not so again Neural Network is required to make random decisions. So because of the decisive nature of Neural Network and the power of Fuzzy system to deal with the uncertainty issue proposed method looks very good for low contrast images REFERENCES [1] Alexey Saenko, Galina Polte and Victor Musalimov Image Enhancement and Image Quality Analysisusing Fuzzy Logic Techniques , 2012 IEEE. [2] Balasubramaniam Jayaram, Kakarla V.V.D.L. Narayana, V. Vetrivel Fuzzy Inference System based Contrast Enhancement EUSFLAT-LFA July, 2011 Aix-les-Bains, France. [3] C.Sasi varnan, A.Jagan,Dr.D.S.Rao Image Quality Assessment Techniques in Spatial Domain IJCST Vol. 2, Issue 3, September 2011 [4] Chuanwei Sun, Hong Liu & Jingao Liu An Image Enhancement Method for Noisy Image , ICALIP 2010 IEEE. [5] G.Maragatham, S.Md Mansoor Roomi, T.Manoj Prabu Contrast Enhancement by object based Histogram Equalization , 2011 IEEE. [6] Khairunnisa Hasikin & Nor Ashidi Mat Isa Enhancement of the low contrast image using fuzzy set theory, 2012 IEEE International Conference on Modeling and Simulation. [7] Ming Zenga,b, Youfu Li b, Qinghao Menga, Ting Yanga, Jian Liua Improving histogrambased image contrast enhancement using graylevel information histogram with application to X-ray images Optik 123 (2012),Elsevier, [8] Milindkumar V. Sarode, Dr. S.A.Ladhake, Dr. Prashant R. Deshmukh Fuzzy system for color image enhancement World Academy of Science, Engineering and Technology, [9] Nafisuddin Khan, K.V. Arya, Manisha Pattanaik An Efficient Image Noise Removal And Enhancement Method , 2010, IEEE. [10] Nachiket Desai, Aritra Chatterjeey, Shaunak Mishra, Dhaval Chudasama, Sunav Choudhary and Sudhirkumar Barai A Fuzzy Logic Based Approach to De-Weather Fog-Degraded Images Sixth International Conference on Computer Graphics, Imaging and Visualization, 2009 IEEE. [11] Qiang Chen, XinXu, QuansenSun, DeshenXia A solution to the deficiencies of image enhancement Elsevier Signal Processing (2010) [12] Raman Maini and Himanshu Aggarwa A Comprehensive Review of Image Enhancement Techniques JOURNAL OF COMPUTING, VOLUME 2, ISSUE 3, MARCH 2010, ISSN [13] R. C. Gonzalez and R. E. Woods. Digital Image Processing. 2nd ed. Prentice Hall, [14] Suzan A. Mahmood Fuzzy Enhancement for Color Image Processing International Conference on Computer Technology and Development, 2009 IEEE. [15] Tom Mélange, Mike Nachtegael,Stefan Schulte, Etienne E. Kerre A fuzzy filter for the removal of random impulse noise in image sequences Elsevier Image and Vision Computing (2011) [16] Xiwen Liu Xiangtan University Xiangtan An Improved Image Enhancement Algorithm Based on Fuzzy Set 2012 International Conference on Medical Physics and Biomedical Engineering Elsevier (2012) [17] Wu Zhihong & Xiao Xiaohong Study on Histogram Equalization International Symposium on Intelligence Information Processing and Trusted Computing,2011 IEEE.